Label construction method, device, electronic device and storage medium
Through business decomposition and feature data analysis, the construction of tags is solved, and the problem of low label construction accuracy in the existing technology is achieved, more accurate business process description and label construction are improved, and the construction accuracy and universality of tags are improved.
Patent Information
- Application Number
- CN202210183402.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-02-25
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2042-02-25
AI Technical Summary
The existing tag construction method is built from top to bottom, and lacks exploration of node relationships per layer, resulting in low label accuracy.
Through business decomposition, obtain the characteristic data of the pending object, determine the business field and business process, perform factor decomposition, extract preset elements to form business objects, perform attribute extraction and transformation rules application, and build labels.
Improves the accuracy of label construction, takes into account the correlation between elements, is universal, and is suitable for different types of data and attributes.
Smart Images

Figure CN114547385B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of label construction, and in particular to a label construction method, device, electronic device and storage medium. Background Art
[0002] With the continuous development of big data and artificial intelligence technologies, more and more companies are undergoing intelligent transformation. Through labeling technology, they mine the characteristics of users, items, and business processes in their businesses, label the characteristics in combination with business language, and form business labels (such as user portrait labels, product labels, etc.) to achieve effective description of the characteristics of business objects. Labels play a very important role in business operations. For example, companies can conduct group marketing for customers based on labels, and make personalized recommendations for users based on product labels and user interest labels. Compared with traditional technologies, the accurate description of the characteristics of business objects based on big data and artificial intelligence technologies greatly improves the efficiency of business operations and reduces operating costs.
[0003] The most common label system construction method is the top-down construction method. Specifically, the label directory tree is planned according to the business field, business process, and business needs, and then the labels are processed in each type of label through statistics, attributes, or model algorithms. This label construction method is only constructed layer by layer from top to bottom, lacking the exploration of the relationship between nodes at each layer, resulting in low accuracy of the constructed labels. Summary of the invention
[0004] The embodiments of the present application provide a label construction method, device, electronic device and storage medium, which improve the label construction accuracy through business decomposition.
[0005] In a first aspect, an embodiment of the present application provides a label construction method, comprising:
[0006] Acquire feature data of the object to be processed;
[0007] Determining the business field of the object to be processed according to the characteristic data of the object to be processed;
[0008] Determining multiple business processes involved in the object to be processed according to the characteristic data of the object to be processed and the business field;
[0009] Decomposing each of the business processes into elements to obtain a plurality of preset elements;
[0010] Selecting some of the preset elements from the multiple preset elements corresponding to each of the business processes to form a business object corresponding to each of the business processes, thereby obtaining multiple business objects corresponding to the multiple business processes;
[0011] Extracting attributes of some of the preset elements included in each of the business objects to obtain multiple attributes of each of the business objects;
[0012] Acquire a transformation rule corresponding to each of the attributes, and perform spatial transformation on the attribute value of each of the attributes according to the transformation rule of each of the attributes to obtain a feature space of each of the business objects;
[0013] According to the feature space of each of the business objects, a label is constructed for the object to be processed.
[0014] In a second aspect, an embodiment of the present application provides a label construction device, including: an acquisition unit and a processing unit;
[0015] The acquisition unit is used to acquire feature data of the object to be processed;
[0016] The processing unit is used to determine the business field of the object to be processed according to the characteristic data of the object to be processed;
[0017] Determining multiple business processes involved in the object to be processed according to the characteristic data of the object to be processed and the business field;
[0018] Decomposing each of the business processes into elements to obtain a plurality of preset elements;
[0019] Selecting some of the preset elements from the multiple preset elements corresponding to each of the business processes to form a business object corresponding to each of the business processes, thereby obtaining multiple business objects corresponding to the multiple business processes;
[0020] Extracting attributes of some of the preset elements included in each of the business objects to obtain multiple attributes of each of the business objects;
[0021] Acquire a transformation rule corresponding to each of the attributes, and perform spatial transformation on the attribute value of each of the attributes according to the transformation rule of each of the attributes to obtain a feature space of each of the business objects;
[0022] According to the feature space of each of the business objects, a label is constructed for the object to be processed.
[0023] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a processor, the processor is connected to a memory, the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device performs the method described in the first aspect.
[0024] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program enables a computer to execute the method described in the first aspect.
[0025] In a fifth aspect, an embodiment of the present application provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer is operable to cause the computer to execute the method described in the first aspect.
[0026] Implementing the embodiments of the present application has the following beneficial effects:
[0027] It can be seen that in the implementation mode of the present application, by analyzing and processing the characteristic data of the object to be processed, a plurality of business processes involved in the object to be processed are obtained, and a plurality of preset elements are formed by decomposing the business process, so as to more accurately characterize the business process. Then, some preset elements are extracted to form a business object, and a label is constructed from the perspective of the business object, that is, a plurality of elements are used as a whole to construct a label. Considering the correlation between the elements, the label is not constructed independently from each element, thereby improving the construction accuracy of the label. Moreover, for the attributes of the elements, they can be transformed into characteristic values through corresponding transformation rules, so no matter what type of data, no matter what attributes it has, its attributes can be transformed into corresponding characteristic values, so the label construction of the present application is universal. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0029] Figure 1 A schematic diagram of a label construction system provided in an embodiment of the present application;
[0030] Figure 2 A flowchart of a label construction method provided in an embodiment of the present application;
[0031] Figure 3 A schematic diagram of a transformation from an attribute space to a feature space provided in an embodiment of the present application;
[0032] Figure 4 A schematic diagram of a label provided in an embodiment of the present application;
[0033] Figure 5 A schematic diagram of a label directory provided in an embodiment of the present application;
[0034] Figure 6 A schematic diagram of another label directory provided in an embodiment of the present application;
[0035] Figure 7 A block diagram of the functional units of a label building device provided in an embodiment of the present application;
[0036] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0037] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0038] The terms "first", "second", "third" and "fourth" etc. in the specification and claims of the present application and the drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.
[0039] Reference to "embodiments" herein means that a particular feature, result, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0040] See also Figure 1 , Figure 1 This is a schematic diagram of a label building system provided in an embodiment of the present application. The label building system includes a first user terminal 10, a label building device 20 and a second user terminal 30. The second user terminal is mainly used for pushing information.
[0041] Exemplarily, the first user terminal 10 may collect feature data of the object to be processed. For example, when the object to be processed is a user, the feature data of the user may be collected through the user's online behavior.
[0042] Furthermore, if the second user terminal 30 has the authority to access the first user terminal 10, it can obtain the feature data of the object to be processed from the first user terminal 10; then determine the business field of the object to be processed based on the feature data of the object to be processed; determine the multiple business processes involved in the object to be processed based on the feature data of the object to be processed and the business field; decompose each business process into elements based on the feature data of the object to be processed to obtain multiple preset elements; select some elements from the multiple preset elements corresponding to each business process to form a business object corresponding to each business process, and obtain multiple business objects corresponding to the multiple business processes; extract attributes from each business object to obtain multiple attributes of each business object; obtain the transformation rule corresponding to each attribute, and perform spatial transformation on each attribute according to the transformation rule of each attribute to obtain the feature space of each business object; and construct a label for the object to be processed based on the feature space of each business object.
[0043] Accordingly, after building the label, the second user terminal 30 can obtain the label of each object to be processed from the label building device 20, determine the information to be pushed of each object to be processed based on the label of each object to be processed, and push the information to be pushed to each object to be processed. For example, based on the label of each object to be processed, the object to be processed that is compatible with a certain product can be determined, and the product information of the product can be pushed to the object to be processed.
[0044] It can be seen that in the implementation mode of the present application, by analyzing and processing the characteristic data of the object to be processed, a plurality of business processes involved in the object to be processed are obtained, and a plurality of preset elements are formed by decomposing the business process, so as to more accurately characterize the business process. Then, some preset elements are extracted to form a business object, and a label is constructed from the perspective of the business object, that is, a plurality of elements are used as a whole to construct a label. Considering the correlation between the elements, the label is not constructed independently from each element, thereby improving the construction accuracy of the label. Moreover, for the attributes of the elements, they can be transformed into characteristic values through corresponding transformation rules, so no matter what type of data, no matter what attributes it has, its attributes can be transformed into corresponding characteristic values, so the label construction of the present application is universal.
[0045] See also Figure 2 , Figure 2 The following is a flow chart of a label construction method provided in an embodiment of the present application. The method is applied to the label construction device 20 described above. The method includes but is not limited to the following steps:
[0046] 201: Obtain feature data of the object to be processed.
[0047] Among them, the objects to be processed in this application include three categories. The first category is people as the objects to be processed, that is, users; the second category is items as the objects to be processed, such as commodities (such as insurance products); the third category is people's behaviors as the objects to be processed, such as users' browsing behaviors and shopping behaviors.
[0048] Exemplarily, when the object to be processed is a user, the feature data is the basic features of the user, such as the user's age, gender, hobbies and social relationships, etc., then the constructed label can be understood as the user's portrait; when the object to be processed is an item, the feature data is the basic features of the item, such as the item's category, color, price, applicable population and sales situation, etc., then the constructed label can be understood as the item's portrait; when the object to be processed is the user's behavior, the feature data is the user's behavioral feature data, such as the number of times the user has shared, the content shared and the browsing history, etc., then the constructed label is the user's behavioral portrait. This application mainly takes the construction of a behavioral portrait based on the user's behavior as an example for explanation, and the construction methods of the other two types of portraits are similar and will not be described again.
[0049] Exemplarily, for different types of objects to be processed, the characteristic data can be obtained in different ways. For example, when the object to be processed is a user, the user database can be accessed through the application program interface to obtain the characteristic data of the user; when the object to be processed is an item, the characteristic data of the item can be obtained by accessing the merchant platform of the item; when the data to be processed is the behavioral characteristic data of the user, the behavioral characteristic data of the user can be obtained by accessing the user terminal of the user.
[0050] 202: Determine the business field of the object to be processed according to the characteristic data of the object to be processed.
[0051] The so-called business domain is a general term for all business processes within a specific business boundary. Therefore, the feature data of the object to be processed can be subjected to entity extraction to obtain at least one entity. For example, for the feature data of "insurance agent shares content with customers", entity extraction can obtain entities "insurance", "agent" and "customer"; then, according to the correspondence between entities and industries, the industry corresponding to each entity is determined. For example, for "insurance", the industry corresponding to the entity is determined to be the insurance industry, for "agent", the industry corresponding to the entity is determined to be the sales industry, and for "customer", the industry corresponding to the entity is determined to be the service industry. Further, the feature data of the object to be processed is subjected to action recognition to obtain the business model involved in the object to be processed, wherein the business model includes an online mode or an offline mode. For example, for the feature data of "insurance agent shares content with customers", its action is identified as "sharing", so the business model corresponding to the feature data is determined to be an online mode. Finally, at least one industry corresponding to the at least one entity and the business model are combined to obtain the business domain of the object to be processed. Continuing with the above example, if at least one of the industries is "insurance industry", "sales industry" or "service industry", and the business model is "online model", then it can be determined that the business field of the object to be processed is "online insurance sales".
[0052] 203: Determine multiple business processes involved in the object to be processed according to the feature data of the object to be processed and the business domain.
[0053] Exemplarily, the feature data of the object to be processed is divided into sentences to obtain multiple sentences; each sentence is parsed to obtain the subject, adverbial, predicate and object in each sentence, wherein the adverbial may include a purpose adverbial, a time adverbial and / or a place adverbial. Continuing with the above example, the subject is "agent", the adverbial is "customer", the predicate is "sharing" and the object is "content". Then, the subject, adverbial, predicate and object in each sentence are combined to obtain the business process corresponding to each sentence. Continuing with the above example, the subject, adverbial, predicate and object are combined to obtain the business process "agent shares content with customers". In addition, online insurance sales for the business field may also include other business processes, for example, it may also include user account registration, user payment order, and so on. Therefore, a business field may include multiple business processes.
[0054] 204: Decompose each of the business processes into elements to obtain a plurality of preset elements.
[0055] Exemplarily, the preset elements of the present application include constraints, goals, subjects, objects, and relationships. Among them, constraints (Constraint) are internal and external restrictions (environment or context) faced by a business process; for example, constraints may include time constraints and / or location constraints, etc., and constraints are generally represented by time adverbs or location adverbs in feature data; goals (Commitment Statement) are descriptions of the goals that need to be achieved in a business process, and are generally represented by purpose adverbs in feature data. Subject (Entity-Actor), a participant ("person" or "system") involved in a business process, is represented by a subject in feature data. Object (Entity-Object), a non-subject entity involved in a business process, is generally represented by an object in feature data; Relationship (Relationship), the relationship between a subject and an object in a business process, is generally represented by a predicate in feature data.
[0056] Therefore, the subject, adverbial (including time adverbial, place adverbial and purpose adverbial), predicate and object of each business process are obtained, so as to decompose each business process into a plurality of preset elements.
[0057] For example, if the feature data is "the insurance agent shared content with the customer in January", the business process corresponding to the feature data is "the agent shared content with the customer in January". The business process is decomposed into elements, and the subject is "agent", the purpose adverbial is "to the customer", the time adverbial is "January", the predicate is "share", and the relationship is "content". Therefore, the subject corresponding to the business process is "agent", the constraint is "January", the goal is "to the customer", and the relationship is "share".
[0058] 205: Selecting some of the preset elements from the multiple preset elements corresponding to each of the business processes to form a business object corresponding to each of the business processes, thereby obtaining multiple business objects corresponding to the multiple business processes.
[0059] Exemplarily, some elements of the present application are subject, object and relationship, that is, subject, object and relationship are selected from the above multiple preset elements to form a business object corresponding to each business process. For multiple business processes, multiple business objects can be obtained.
[0060] For example, if the business process is that an agent shares content with a customer, the obtained business object is {agent, customer, sharing}.
[0061] 206: Extract attributes of some of the preset elements included in each of the business objects to obtain multiple attributes of each of the business objects.
[0062] Exemplarily, attributes are extracted from each preset element in the business object corresponding to each business process to obtain attributes corresponding to each preset element, wherein each preset element corresponds to one or more attributes; then, all attributes corresponding to this part of the preset elements are merged to obtain multiple attributes corresponding to the business object.
[0063] For example, for the business object: {agent, customer, sharing}, the attributes of the preset element "subject", that is, the agent, are extracted, and its attributes include age, gender, company length of service, personal honors, level, etc.; the attributes of the preset element "object", that is, the customer, are extracted, and its attributes include age, gender, etc.; the attributes of the preset element "relationship", that is, sharing, are extracted, and its attributes include sharing failure, sharing success, and probability preference of sharing content.
[0064] 207: Acquire a transformation rule corresponding to each of the attributes, and perform spatial transformation on each of the attributes according to the transformation rule of each of the attributes to obtain a feature space of each of the business objects.
[0065] Exemplarily, a transformation rule corresponding to each attribute is selected from a plurality of preset transformation rules. Optionally, the plurality of preset transformation rules in the present application mainly include three categories, specifically: mapping, statistics and model prediction. Among them, statistics use statistical analysis methods to transform attribute values; for example, summing, averaging, finding the maximum value, finding the minimum value, etc. of order amounts; mapping is to transform attribute values using mapping rules; for example, mapping the age expressed in text to age groups corresponding to teenagers, young people, middle-aged people, and the elderly; model prediction is to transform attribute values using a trained model; for example, probabilistically predicting the user's preference for sharing content, and obtaining the probability distribution characteristics of the user's content sharing preference, that is, predicting the probability of users sharing various types of content through the model.
[0066] Specifically, according to the characteristic data of the object to be processed, the attribute value of each attribute is obtained; if the attribute value of each attribute is represented by text, the transformation rule corresponding to each attribute is determined to be mapping, that is, the attribute value expressed by fuzzy concepts, and its corresponding change rule is mapping. For example, the age of the agent is a teenager, and the corresponding transformation rule is mapping, and the teenager needs to be mapped to a specific age group; if the attribute value of each attribute is represented by numbers, the transformation rule corresponding to each attribute is determined to be statistics. For example, the number of times the agent shares content, the total number of times the agent shares content within a preset time period can be statistically counted; if the attribute value of each attribute is represented by probability, the transformation rule corresponding to each attribute is determined to be model prediction. For example, for the agent's content sharing, its attribute value has a sharing preference characteristic distribution, and the characteristic distribution needs to be represented by probability, and the probability of the agent sharing various types of content is predicted by the model.
[0067] Therefore, after obtaining the transformation rule corresponding to each attribute, the attribute value of each attribute is transformed based on the transformation rule corresponding to each attribute to obtain the feature value corresponding to each attribute (i.e., the text, value, or character, etc. describing each attribute). Specifically, if the transformation rule of each attribute is mapping, mapping is performed according to the attribute value of each attribute to obtain the feature value corresponding to each attribute; if the transformation rule of each attribute is statistics, the number of times the attribute value of each attribute appears within a preset time period is counted, and the number is used as the feature value corresponding to each attribute; if the transformation rule of each attribute is model prediction, the trained prediction model and the attribute value of each attribute are used to predict the probability distribution under each attribute, and the probability distribution is used as the feature value corresponding to each attribute.
[0068] Finally, the feature values corresponding to each attribute are combined to obtain the feature space corresponding to each business object.
[0069] Generally speaking, a feature value corresponding to an attribute is obtained by transformation through mapping or statistics, and multiple feature values corresponding to the attribute are obtained by transformation through model prediction.
[0070] For example, still using the above example of an agent sharing content with a customer, by transforming the attribute value corresponding to each attribute, we can get the following: Figure 3 The feature space shown.
[0071] 208: Construct a label for the object to be processed according to the feature space of each of the business objects.
[0072] For example, Figure 4As shown, the characteristic value corresponding to each attribute is used as a label of the object to be processed to obtain the label of the object to be processed.
[0073] It can be seen that in the implementation mode of the present application, by analyzing and processing the characteristic data of the object to be processed, a plurality of business processes involved in the object to be processed are obtained, and a plurality of preset elements are formed by decomposing the business process, so as to more accurately characterize the business process. Then, some preset elements are extracted to form a business object, and a label is constructed from the perspective of the business object, that is, a plurality of elements are used as a whole to construct a label. Considering the correlation between the elements, the label is not constructed independently from each element, thereby improving the construction accuracy of the label. Moreover, for the attributes of the elements, they can be transformed into characteristic values through corresponding transformation rules, so no matter what type of data, no matter what attributes it has, its attributes can be transformed into corresponding characteristic values, so the label construction of the present application is universal.
[0074] In one embodiment of the present application, after building a label for the object to be processed, the method further includes:
[0075] Construct a tree structure, wherein the tree structure includes a parent node, multiple first-level child nodes and multiple second-level child nodes, wherein the number of the multiple first-level child nodes is determined by the number of all attributes corresponding to the multiple business objects, and the number of the multiple second-level child nodes is determined by the number of all feature values corresponding to all attributes corresponding to the multiple business objects. Specifically, Figure 5 As shown, a tree structure is constructed, and multiple first-level child nodes are constructed in the tree structure according to the number of attributes, and multiple second-level child nodes are constructed according to the number of characteristic values corresponding to each attribute; further, as Figure 5 As shown, the business domain of the business object is mounted on the parent node to obtain a first-level directory; multiple attributes of each business object are mounted on the multiple first-level nodes in sequence to obtain a second-level directory; the characteristic values corresponding to each attribute of the multiple attributes of each business object are mounted on the multiple second-level child nodes in sequence to obtain a third-level directory; the tree structure mounted with the first-level directory, the second-level directory and the third-level directory is used as the directory of the label.
[0076] It should be noted that the above-mentioned construction of the label directory is mainly to form a label directory by classification management from the perspective of the business process goal. Generally speaking, the business domain of the business object is taken as the top-level directory (i.e., the first-level directory), and then, according to the goal of the business process corresponding to the business object, the sub-nodes of each layer are designed layer by layer; until the last layer of sub-nodes, that is, each attribute is taken as a first-level sub-node, and then the characteristic values corresponding to each attribute are mounted on the second-level nodes corresponding to each attribute to form the label directory.
[0077] It can be seen that by constructing a directory for the labels of the objects to be processed, the constructed labels can be added with the business field and the various sub-fields (i.e., various attributes) related to the business field, so as to facilitate the adaptation of the labels of the objects to be processed to the business field and provide the business field with the label information of the user.
[0078] In one embodiment of the present application, after building the directory of tags, the method further includes:
[0079] Obtain a target attribute from among the multiple attributes of each of the business objects, wherein the target attribute is determined by the subject, predicate, object, and adverbial in the feature data of the object to be processed, that is, the attributes related to the identity in the multiple attributes corresponding to the subject, the attributes related to the identity in the multiple attributes corresponding to the object, the attributes related to the identification of the action in the predicate, and the attributes related to the identification of the purpose in the adverbial are all taken as the target attribute; obtain at least one identity identifier corresponding to the target attribute of each of the business objects; associate at least one identity identifier corresponding to the target attribute of each of the business objects to obtain a target identity identifier of the target attribute of each of the business objects. And take the target identity identifier of the target attribute of each of the business objects as a new feature value of the target attribute.
[0080] Furthermore, if Figure 6 As shown, since the target identity is a new characteristic value, the original tree structure does not construct a third-level node corresponding to the characteristic value. Therefore, a third-level node is newly added for the target attribute of each business object in the directory of the label, and the target identity of the target attribute is mounted on the newly added third-level node, and the target identity (characteristic value) is mounted on the third-level node corresponding to the target attribute in the tree structure.
[0081] The following uses the identity of the associated subject as an example to illustrate the process of obtaining the target identity.
[0082] For example, if the subject is a user, and at least one identity identifier of the user includes the user's WeChat ID, APP registration ID, and mobile phone number; then, the user's WeChat ID, APP registration ID, and mobile phone number are associated to obtain the target identity identifier of the subject, that is, the user's multiple identity identifiers are spliced (i.e. associated) to obtain the target identity identifier of {WeChat ID, APP registration ID, mobile phone number}.
[0083] It can be seen that in this implementation, by associating the identity identifier, the uniqueness of the identity identifier is guaranteed, which facilitates the accuracy of the feature value of each node to be obtained later.
[0084] See also Figure 7 , Figure 7 The functional unit composition block diagram of a label building device provided in an embodiment of the present application. The label building device 700 includes: an acquisition unit 701 and a processing unit 702;
[0085] An acquisition unit 701 is used to acquire feature data of an object to be processed;
[0086] The processing unit 702 is used to determine the business field of the object to be processed according to the characteristic data of the object to be processed;
[0087] Determining multiple business processes involved in the object to be processed according to the characteristic data of the object to be processed and the business field;
[0088] Decomposing each of the business processes into elements to obtain a plurality of preset elements;
[0089] Selecting some of the preset elements from the multiple preset elements corresponding to each of the business processes to form a business object corresponding to each of the business processes, thereby obtaining multiple business objects corresponding to the multiple business processes;
[0090] Extracting attributes of some of the preset elements included in each of the business objects to obtain multiple attributes of each of the business objects;
[0091] Acquire a transformation rule corresponding to each of the attributes, and perform spatial transformation on the attribute value of each of the attributes according to the transformation rule of each of the attributes to obtain a feature space of each of the business objects;
[0092] According to the feature space of each of the business objects, a label is constructed for the object to be processed.
[0093] In one embodiment of the present application, in terms of determining the business field of the object to be processed according to the feature data of the object to be processed, the processing unit 702 is specifically configured to:
[0094] Performing entity extraction on the feature data of the object to be processed to obtain at least one entity;
[0095] According to the corresponding relationship between industries and entities, determine the industry corresponding to each of the entities;
[0096] Performing action recognition on the feature data of the object to be processed to obtain a business model involved in the object to be processed, wherein the business model includes an online mode or an offline mode;
[0097] At least one industry corresponding to the at least one entity and the business model are combined to obtain the business field of the object to be processed.
[0098] In one embodiment of the present application, in determining multiple business processes involved in the object to be processed according to the feature data of the object to be processed and the business field, the processing unit 702 is specifically configured to:
[0099] Segmenting the characteristic data of the object to be processed to obtain a plurality of sentences;
[0100] Performing grammatical analysis on each of the sentences to obtain the subject, adverbial, predicate and object in each of the sentences;
[0101] Combining the subject, adverbial, predicate and object in each of the sentences to obtain the business process corresponding to each of the sentences;
[0102] The multiple business processes corresponding to the multiple statements are combined to obtain the multiple business processes.
[0103] In one embodiment of the present application, in terms of obtaining the transformation rule corresponding to each of the attributes, the processing unit 702 is specifically configured to:
[0104] Acquire the attribute value of each attribute according to the feature data of the object to be processed;
[0105] If the attribute value of each of the attributes is represented by text, determining the transformation rule corresponding to each of the attributes as a mapping;
[0106] If the attribute value of each of the attributes is represented by a number, determining that the transformation rule corresponding to each of the attributes is statistical;
[0107] If the attribute value of each of the attributes is represented by probability, then the transformation rule corresponding to each of the attributes is determined as model prediction.
[0108] In one embodiment of the present application, in acquiring the transformation rule corresponding to each of the attributes, performing spatial transformation on each of the attributes according to the transformation rule of each of the attributes, and obtaining the feature space of each of the business objects, the processing unit 702 is specifically configured to:
[0109] If the transformation rule of each attribute is mapping, mapping is performed according to the attribute value of each attribute to obtain the feature value corresponding to each attribute;
[0110] If the transformation rule of each attribute is statistics, the number of times the attribute value of each attribute appears within a preset time period is counted, and the number is used as the feature value corresponding to each attribute;
[0111] If the transformation rule of each attribute is model prediction, the trained prediction model and the attribute value of each attribute are used to predict the probability distribution of each attribute, and the probability distribution is used as the feature value corresponding to each attribute;
[0112] The feature values corresponding to each of the attributes are combined to obtain a feature space of each of the business objects.
[0113] In one embodiment of the present application, the processing unit 702 is further configured to construct a tree structure, wherein the tree structure includes a parent node, a plurality of first-level child nodes, and a plurality of second-level child nodes, wherein the number of the plurality of first-level child nodes is determined by the number of all attributes corresponding to the plurality of business objects, and the number of the plurality of second-level child nodes is determined by the number of all feature values corresponding to all attributes corresponding to the plurality of business objects;
[0114] Mount the business domain of the business object on the parent node to obtain a first-level directory;
[0115] Mounting the multiple attributes of each of the business objects in sequence on the multiple first-level nodes to obtain a second-level directory;
[0116] Mounting the characteristic value corresponding to each of the multiple attributes of each of the business objects on the multiple secondary child nodes in sequence to obtain a third-level directory;
[0117] The tree structure in which the first-level directory, the second-level directory, and the third-level directory are mounted is used as the directory of the label.
[0118] In one embodiment of the present application, the processing unit 702 is further configured to:
[0119] Obtain at least one identity corresponding to the target attribute of each of the business objects;
[0120] Associating at least one identity identifier corresponding to the target attribute of each of the business objects to obtain a target identity identifier for the target attribute of each of the business objects;
[0121] A third-level node is newly added in the directory of the tag for the target attribute of each business object, and the target identity of the target attribute is mounted on the newly added third-level node.
[0122] See also Figure 8 , Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 7As shown, the electronic device 800 includes a transceiver 801, a processor 802 and a memory 803. They are connected via a bus 804. The memory 803 is used to store computer programs and data, and can transmit the data stored in the memory 803 to the processor 802.
[0123] The processor 802 is used to read the computer program in the memory 803 and perform the following operations:
[0124] Controlling the transceiver 801 to obtain feature data of the object to be processed;
[0125] Determining the business field of the object to be processed according to the characteristic data of the object to be processed;
[0126] Determining multiple business processes involved in the object to be processed according to the characteristic data of the object to be processed and the business field;
[0127] Decomposing each of the business processes into elements to obtain a plurality of preset elements;
[0128] Selecting some of the preset elements from the multiple preset elements corresponding to each of the business processes to form a business object corresponding to each of the business processes, thereby obtaining multiple business objects corresponding to the multiple business processes;
[0129] Extracting attributes of some of the preset elements included in each of the business objects to obtain multiple attributes of each of the business objects;
[0130] Acquire a transformation rule corresponding to each of the attributes, and perform spatial transformation on the attribute value of each of the attributes according to the transformation rule of each of the attributes to obtain a feature space of each of the business objects;
[0131] According to the feature space of each of the business objects, a label is constructed for the object to be processed.
[0132] In one implementation of the present application, in terms of determining the business field of the object to be processed according to the feature data of the object to be processed, the processor 802 is specifically configured to perform the following steps:
[0133] Performing entity extraction on the feature data of the object to be processed to obtain at least one entity;
[0134] According to the corresponding relationship between industries and entities, determine the industry corresponding to each of the entities;
[0135] Performing action recognition on the feature data of the object to be processed to obtain a business model involved in the object to be processed, wherein the business model includes an online mode or an offline mode;
[0136] At least one industry corresponding to the at least one entity and the business model are combined to obtain the business field of the object to be processed.
[0137] In one embodiment of the present application, in determining the multiple business processes involved in the object to be processed according to the feature data of the object to be processed and the business field, the processor 802 is specifically configured to perform the following steps:
[0138] Segmenting the characteristic data of the object to be processed to obtain a plurality of sentences;
[0139] Performing grammatical analysis on each of the sentences to obtain the subject, adverbial, predicate and object in each of the sentences;
[0140] Combining the subject, adverbial, predicate and object in each of the sentences to obtain the business process corresponding to each of the sentences;
[0141] The multiple business processes corresponding to the multiple statements are combined to obtain the multiple business processes.
[0142] In one embodiment of the present application, in terms of obtaining the transformation rule corresponding to each attribute, the processor 802 is specifically configured to perform the following steps:
[0143] Acquire the attribute value of each attribute according to the feature data of the object to be processed;
[0144] If the attribute value of each of the attributes is represented by text, determining the transformation rule corresponding to each of the attributes as a mapping;
[0145] If the attribute value of each of the attributes is represented by a number, determining that the transformation rule corresponding to each of the attributes is statistical;
[0146] If the attribute value of each of the attributes is represented by probability, then the transformation rule corresponding to each of the attributes is determined as model prediction.
[0147] In one embodiment of the present application, in terms of obtaining the transformation rule corresponding to each of the attributes, performing spatial transformation on each of the attributes according to the transformation rule of each of the attributes, and obtaining the feature space of each of the business objects, the processor 802 is specifically configured to perform the following steps:
[0148] If the transformation rule of each attribute is mapping, mapping is performed according to the attribute value of each attribute to obtain the feature value corresponding to each attribute;
[0149] If the transformation rule of each attribute is statistics, the number of times the attribute value of each attribute appears within a preset time period is counted, and the number is used as the feature value corresponding to each attribute;
[0150] If the transformation rule of each attribute is model prediction, the trained prediction model and the attribute value of each attribute are used to predict the probability distribution of each attribute, and the probability distribution is used as the feature value corresponding to each attribute;
[0151] The feature values corresponding to each of the attributes are combined to obtain a feature space of each of the business objects.
[0152] In one embodiment of the present application, the processor 802 is further configured to perform the following steps:
[0153] Constructing a tree structure, wherein the tree structure includes a parent node, a plurality of first-level child nodes, and a plurality of second-level child nodes, wherein the number of the plurality of first-level child nodes is determined by the number of all attributes corresponding to the plurality of business objects, and the number of the plurality of second-level child nodes is determined by the number of all feature values corresponding to all attributes corresponding to the plurality of business objects;
[0154] Mount the business domain of the business object on the parent node to obtain a first-level directory;
[0155] Mounting the multiple attributes of each of the business objects in sequence on the multiple first-level nodes to obtain a second-level directory;
[0156] Mounting the characteristic value corresponding to each of the multiple attributes of each of the business objects on the multiple secondary child nodes in sequence to obtain a third-level directory;
[0157] The tree structure in which the first-level directory, the second-level directory, and the third-level directory are mounted is used as the directory of the label.
[0158] In one embodiment of the present application, the processor 802 is further configured to perform the following steps:
[0159] Obtain at least one identity corresponding to the target attribute of each of the business objects;
[0160] Associating at least one identity identifier corresponding to the target attribute of each of the business objects to obtain a target identity identifier for the target attribute of each of the business objects;
[0161] A third-level node is newly added in the directory of the tag for the target attribute of each business object, and the target identity of the target attribute is mounted on the newly added third-level node.
[0162] Specifically, the transceiver 801 may be Figure 7 The acquisition unit 701 of the label construction device 700 of the embodiment described above, the processor 802 may be Figure 7The processing unit 702 of the label construction device 700 of the embodiment described.
[0163] It should be understood that the electronic devices in this application may include smart phones (such as Android phones, iOS phones, Windows Phone phones, etc.), tablet computers, PDAs, laptops, mobile Internet devices MID (Mobile Internet Devices, referred to as: MID) or wearable devices, etc. The above electronic devices are only examples, not exhaustive, and include but are not limited to the above electronic devices. In practical applications, the above electronic devices may also include: smart vehicle terminals, computer equipment, etc.
[0164] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement part or all of the steps of any one of the label construction methods recorded in the above method embodiments.
[0165] An embodiment of the present application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute part or all of the steps of any one of the label construction methods recorded in the above method embodiments.
[0166] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.
[0167] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0168] In the several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the device or unit can be electrical or other forms.
[0169] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0170] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of a software program module.
[0171] If the integrated unit is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or CD-ROM and other media that can store program codes.
[0172] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, and the memory can include: a flash drive, a read-only memory (English: Read-Only Memory, abbreviated as: ROM), a random access memory (English: Random Access Memory, abbreviated as: RAM), a magnetic disk or an optical disk, etc.
[0173] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of the present application. At the same time, for general technical personnel in this field, according to the idea of the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
Claims
1. A label construction method, characterized in that: include: Acquire feature data of the object to be processed; Determining the business field of the object to be processed according to the characteristic data of the object to be processed; Determine multiple business processes involved in the object to be processed according to the feature data of the object to be processed and the business field; including: segmenting the feature data of the object to be processed to obtain multiple sentences; performing grammatical analysis on each of the sentences to obtain the subject, adverbial, predicate and object in each of the sentences; combining the subject, adverbial, predicate and object in each of the sentences to obtain the business process corresponding to each of the sentences; combining the multiple business processes corresponding to the multiple sentences to obtain the multiple business processes; Decomposing each of the business processes into elements to obtain a plurality of preset elements; Selecting some of the preset elements from the multiple preset elements corresponding to each of the business processes to form a business object corresponding to each of the business processes, thereby obtaining multiple business objects corresponding to the multiple business processes; Extracting attributes of some of the preset elements included in each of the business objects to obtain multiple attributes of each of the business objects; Acquiring a transformation rule corresponding to each of the attributes, and performing spatial transformation on the attribute value of each of the attributes according to the transformation rule of each of the attributes to obtain a feature space of each of the business objects; including: Acquire the attribute value of each attribute according to the feature data of the object to be processed; If the attribute value of each attribute is represented by text, the transformation rule corresponding to each attribute is determined to be mapping, and mapping is performed according to the attribute value of each attribute to obtain the feature value corresponding to each attribute; if the attribute value of each attribute is represented by numbers, the transformation rule corresponding to each attribute is determined to be statistics, and the number of occurrences of the attribute value of each attribute within a preset time period is counted, and the number is used as the feature value corresponding to each attribute; if the attribute value of each attribute is represented by probability, the transformation rule corresponding to each attribute is determined to be model prediction, and the trained prediction model and the attribute value of each attribute are used to predict the probability distribution under each attribute, and the probability distribution is used as the feature value corresponding to each attribute; Combining the feature values corresponding to each of the attributes to obtain a feature space of each of the business objects; According to the feature space of each of the business objects, a label is constructed for the object to be processed.
2. The method according to claim 1, characterized in that The determining, based on the characteristic data of the object to be processed, the business field of the object to be processed includes: Performing entity extraction on the feature data of the object to be processed to obtain at least one entity; According to the corresponding relationship between industries and entities, determine the industry corresponding to each of the entities; Performing action recognition on the feature data of the object to be processed to obtain a business model involved in the object to be processed, wherein the business model includes an online mode or an offline mode; At least one industry corresponding to the at least one entity and the business model are combined to obtain the business field of the object to be processed.
3. The method according to claim 1, characterized in that The method further comprises: Constructing a tree structure, wherein the tree structure includes a parent node, a plurality of first-level child nodes, and a plurality of second-level child nodes, wherein the number of the plurality of first-level child nodes is determined by the number of all attributes corresponding to the plurality of business objects, and the number of the plurality of second-level child nodes is determined by the number of all feature values corresponding to all attributes corresponding to the plurality of business objects; Mount the business domain of the business object on the parent node to obtain a first-level directory; Mounting the multiple attributes of each of the business objects in sequence on the multiple first-level child nodes to obtain a second-level directory; Mounting the characteristic value corresponding to each of the multiple attributes of each of the business objects on the multiple secondary child nodes in sequence to obtain a third-level directory; The tree structure in which the first-level directory, the second-level directory, and the third-level directory are mounted is used as the directory of the label.
4. The method according to claim 3, characterized in that The method further comprises: Acquire a target attribute from among the multiple attributes of each of the business objects, wherein the target attribute is determined by a subject, a predicate, an object, and an adverbial in the feature data of the object to be processed; Obtain at least one identity corresponding to the target attribute of each of the business objects; Associating at least one identity identifier corresponding to the target attribute of each of the business objects to obtain a target identity identifier for the target attribute of each of the business objects; A third-level node is newly added in the directory of the tag for the target attribute of each business object, and the target identity of the target attribute is mounted on the newly added third-level node.
5. A label construction device, characterized in that: The device is used to perform the method according to any one of claims 1 to 4, the method comprising: an acquisition unit and a processing unit; The acquisition unit is used to acquire feature data of the object to be processed; The processing unit is used to determine the business field of the object to be processed according to the characteristic data of the object to be processed; Determining multiple business processes involved in the object to be processed according to the characteristic data of the object to be processed and the business field; Decomposing each of the business processes into elements to obtain a plurality of preset elements; Selecting some of the preset elements from the multiple preset elements corresponding to each of the business processes to form a business object corresponding to each of the business processes, thereby obtaining multiple business objects corresponding to the multiple business processes; Extracting attributes of some of the preset elements included in each of the business objects to obtain multiple attributes of each of the business objects; Acquire a transformation rule corresponding to each of the attributes, and perform spatial transformation on the attribute value of each of the attributes according to the transformation rule of each of the attributes to obtain a feature space of each of the business objects; According to the feature space of each of the business objects, a label is constructed for the object to be processed.
6. An electronic device, characterized in that: include: A processor and a memory, wherein the processor is connected to the memory, the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device executes the method as described in any one of claims 1 to 4.
7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method according to any one of claims 1 to 4.
